Amit Saxena, ServiceNow and Steven Lock-Campbell, U.S. Department of Agriculture | ServiceNow Knowledge 23

This is Techstrong tv. Hello and welcome back to ServiceNow Knowledge 2023. We're here with Ahed Zena from ServiceNow, and we've got Steven Lock Campbell from the U S D A.

We're talking about all things automation. Ahed, welcome to show. Welcome, always pleasure to be here.

People are throwing around the buzz phrase, hyper automation. We've been doing automation since, I don't know, they put mainframes in buildings in the first floor of New York City. Do automate insurance forms in the fifties.

What's changing, what's different? What's hyper automation about? Well, other than that, buzzword, uh, there's quite a few things have changed, Mike.

So first around this, decades as, as you're talking about from IBM mainframe days, all of these technologies does have a purpose to try to automate as much as possible, but they were purpose built on a specific set of use case. So they get start misused for all variety of things. Now, as you fast forward, you start looking at a lot of AI-based technology with IDP document capture, uh, RPAs.

Now generative ai, what do you need is a toolbox to be able to pick right technology for the right use case. So that's one of the paramount pillar of hyper automation. The second thing which hyper automation indicates is speed of innovation.

We are not in the time where it's going to take you weeks and months and years to be implementing your automation projects. Business users are demanding three to four weeks from the time they get your request. So how can you move faster?

And that brings up all this tech tooling we are talking about seamlessly automatically integrated with pre-built assets and use of AI to deliver that. And the third thing, which has changed drastically all of this needs in the last four to five years is our experience, customer experience, supplier experience, partner experience of whatever you call the end users are demanding a better experience for whatever they touch and use. So the short answer is hyper automation is bringing all these three pillars together and there is no better time in the word than today.

All right, Steven. Government agencies, the one thing we all know about 'em, they have processes and lots of paperwork we Do. And the other thing that we're we're experts at is, is more with less, right?

So we've got the same amount of workload, if not more, with a third or less of the staff. So we have to move faster and uh, we have to, you know, be able to, um, really bridge gaps where they exist and uh, you know, we have to cross cross those bridges. Are there particular areas of automation that you guys are prioritizing?

What's kind of top of mind for you? So for me, right now I'm working on a project to eliminate duplicative data entry. Uh, it's pretty low hanging fruit.

I know. Um, you know, it's not super sexy. It's not uh, uh, saving trillions of dollars, but it makes, uh, a couple of employees' lives better.

It eliminates errors in data entry. It eliminates the data entry itself. Um, it, you know, an RPA bot doesn't take vacation time.

It doesn't have to worry that there are three Mondays this year that are holidays that start pay periods, right? So, so what we're working on is, um, is for payroll and you know, we don't have to worry about those employee or human based things like mistakes like vacation time, like life happening to a, to a robot. How are we going about that?

Because this is a pet peeve of mine forever. I mean, you go to the doctor's office and I'm entering the same data again and again and again. Every time I go there, everywhere I go, I gotta enter the same data over and over again.

I just did That last energy Shots and they're telling me that they know me and I'm like, Not really. So, um, what is the challenge? Cuz we've been at this for a while and from your perspective, when you talk to other customers, what is the issue?

We talk about digital transformation all the time. We know we have these paper-based processes. One wise person one said there'll be a paperless bathroom before there'll be a paperless office.

They're probably still right, but, um, what are you seeing as the hurdle? Well, um, there are certain industries who are able to take on and move faster. Now finance, even though we can complain about palm complex processes, they are cutting edge and deploying all these technologies.

Uh, nowadays your loans can get approved in under 90 seconds. I was vacationing in Hawaii and lost my credit card's. Real story mic, right?

And here I am, traditional calling to a credit card company. I've lost it. So because I don't want it getting misused, right?

Bank of Amer, uh, I'm American Express here to say, no worries. Where are you in Hawaii? We can overnight the car to you, alright, from their systems.

Okay, I love it. But what magical thing happened after that, in 30 minutes I get a text saying that your Apple watch digital wallet has been updated with the new card. I don't even know the number in 10 minutes I had a card, which I can go and use it, right?

That's hyper automation. Now in the, now in the case of these medical, uh, and life sciences and insurance, things are moving slow because of a lot of it is compliance related. Some of this is even the digital leaders buying that vision, which finance leaders have bought in that my goal, I don't want to have my nurse spending 30% of the time doing a patient transfer from one to other patient.

Nurses do not join to be doing the patient transfer paperwork. And once they realize that there are better automation technologies available in which we have demonstrated it, they soon realize that employee satisfaction score going up, they have more time available, they can take care of their patients better, right? And they don't have to pick up a phone just to check if it validates the insurance claim, whether it's going to take it, which is available, which is not available.

So I think it's not a one size fit all approach. Certain industries are moving faster and all, um, and that's where the maturity curve of digital transformation comes in. And we have some work to do.

The perception is that employees are resistant to change. But what do you see when you bring in automation? What's the culture like?

What's the general vibe you get from the people you're working with? So once we're able to actually articulate that, you know, a machine isn't replacing your job, it's replacing the parts of your job that we all hate that are critically important, like payroll getting paid is, I find it to be very important. Um, but we're able to focus that employee on higher value tasks.

Uh, to his point, you know, the reason that they joined wasn't to take data from one system and put it into another, especially in government. It was to make a difference, right? That's why we're all here is to, is to make the United States government especially, uh, in the department run better, be better and more efficient.

And once you identify that and explain that our, our staff has been really receptive. They're like, wait, I can actually do my job instead of all of the paperwork. Uh, you need More people like Steve to get it and they unlock this value and then it just goes viral.

We've been talking about AI all week. Yes. What should people expect from AI as it applies to automation?

I mean, it could be the next big kicker for innovation, but what's real and what's, you know, theoretical, Right? We've made several announcements this morning, Mike, but let's talk about two or three things which really makes the automation and, and making all of our product and engineering guys super excited If part of building any automation is the developer tooling the time somebody has to spend time in building. So that's where we can use generative AI right at the get go.

Like next best action using a large scale model, right? Using a preexisting, uh, training sets. So that will speed up your time for development.

Second thing is around, um, the requirements gathering process. Many times in order to understand what you are automating somebody's describe you, now we can take that text input, convert that text to flow text to app directly with the advanced algorithms, which are available right now instead of you drawing a old classic visual chart or drawing the three page requirement document you can interact to in indicate your requirements. Once your requirements are understood, your end automation is gonna be much more robust, faster, and quicker.

So that's one side of it developer tooling. The second side of it is once we unlock certain AI services to the end users, just like in the past classic example is fraud detection or things like that, you can now have similar services like content summarizations of all your cases, generic prompt for q and a, your AI search, our AI search or other searches does a pretty good job. They give you three different articles, but they still expect you to read those three articles.

Now I can take those three articles, which was came in, which was relevant, send it to a gen AI for summarization and get a one page summary of it. So that saved a tremendous time for me. Again, going back to what, see when you were talking about can I first of all be relevant mm-hmm.

Highly accurate and saves me nine extra minutes of time of getting that data. So happy employees, more accuracy. So that kind of things I'm more excited about because we as product people, we want to unlock multiple use cases across customer workflows, employee workflows, and create a workflows so that they can use it for the use cases they want to use it for.

I'm not entirely sure about that use case cuz what I heard is I should just write the summary and skip the article. You can do so as well if you, if you have a teleprompter written down, oh, sorry, a prompt written down for exactly asking for that, but I am sure you can look, um, it's gonna happen someday. Mm-hmm.

Is the approach to this therefore that I should just kind of make a list of the 10 things I hate about my job and then we can start knocking 'em off through automation. I mean, is that a viable way to go? It's a great way to start a conversation.

You know, I'm not a, an expert in all of the things for which I have oversight, right? Or I'm not an expert in the nuanced process of of each workflow, uh, that my division, you know, has. But what I do know is I can't elicit that information.

I can't generate those requirements without actually engaging with staff who have done this for their entire career. So I think that, you know, it's, it's, it's, uh, an evolution and you're never really done, but you have to have somewhere to start a conversation. Now, one of the things that's really important is just because you can use a tool like RPA doesn't mean that you should.

Um, and actually on my project, we had a few different tools in the toolbox and we struck them out for various reasons. And we landed on RPA because, you know, it met requirements in a way that nothing else really did. And I think that that's to your point about having a full toolkit, full toolkit is it's RPA especially isn't the end all be all to everything.

AI isn't going to be the end all be all to everything, but if we can use a little bit of everything to get it all better, then we've made that progress. We've helped that employee and honestly asking the employees what you hate about your job, take it with a grain of salt, of course. But it's a great place to start that conversation of what are we doing?

Why do our processes not meet what our employees expect? Right? You might not get the answer you're looking for because you know, they may Yes, Yes.

You have to put the right prompt around it, as you said Before. There you go. We have had automation forever in a day, and as far as I can tell, we have now millions of islands of automation.

So do we need to kind of think this through a little bit and say, how are we gonna stitch all these integrations and automations together in a way that creates something that's cohesive? What's the next step for where You're talking about the technologies or are you talking about customers experience? Both the, The customers and the platforms?

If you look around them, they have all kinds of things that they have automated in isolation from everything else, but it feels like we need to, as part of some digital transformation, integrate and unify all that. So how do we get there? I think that comes back to the philosophy we've been trying to bring out in the market from ServiceNow with without hyper automation.

Now it's, it, it gets shielded most of the time because of a lot of noise in the government around, um, in the, and and the nature of this word, which everybody understands is differently the moment you start throwing automation fabric and hyper automation. But one important principle from ServiceNow, as we think about hyper automation or anybody who want to look at us as a, uh, credible vendor, is we have truly taken a time to deliver all this tooling kits with Steve. You were talking about ai, RPA integration all on a single platform.

And what that means is truly all of these technologies can exchange the data seamlessly. You don't have to integrate the automation technologies themselves. The second part is, it's some are, some of it is future proof.

Mike, to your point, so four months ago or five months ago, we had no direct plans of launching Janai to the level what we are doing right now. Many companies did not have it, but because of the solid nature of our architecture, of our platforms and all, guess what? The speed of innovation we are launching.

We have launched tons of gen AI initiatives. In fact, every single product team in ServiceNow is working on it. And we are bringing it to the market May 4th.

We already had a gen AI controller. Now guess what, what customer has to do, cust none of our customers have to upgrade anything. Their investment in ServiceNow, they are realizing the value.

So that's the difference we are seeing with eliminating the islands of automations. And now that doesn't mean that you need a only single team going forward because that's not ible. You don't wanna be the unpopular guy saying that, well, let's bring all the automations into a single team.

Nobody's gonna buy them. You want to still continue doing decentralized teams. However, the difference is guys, when you set up your decentralized team, use a single platform so you can reuse, discover, and have the right set of governance, but let them move at their own speed, right?

That's the difference. If I could just pick up a little bit, I, I think having a vision and a holistic approach is really important. You have to have leadership in place and, and that empower staff to say, okay, what's next?

We've done all of these things in our silos, which we spent time knocking down, right? And we've re siloed. We have to knock 'em down again to an extent and think about the way that we can, okay, now that we're all in the same pool and the, the commonality of the platform is a conversation that I have weekly is, you know, I already have access to this data.

Let me use X, Y, and Z. Um, so I think having a, a leadership vision that says, okay, we did this. What's next?

How do we keep optimizing? How do we keep improving? Um, and then you get to, to use the cool new tools that ServiceNow is putting out in the marketplace.

What do you know now that you kind of wish you knew a year ago, maybe two years ago when you first started this whole thing? I wish that I knew how well it was all gonna work. Um, when we first started talking about rpa, we explored a couple of different vendors and there was just some components and we'd heard rumors that there was a, there was ServiceNow, r p a in the pipeline.

And I wish that I would've known that it was going to be as robust and the, that the design studio was gonna be honestly as fun as it is to, to use. Um, because we probably would've put a pin in a couple of things and said, we're gonna solve this in a really amazing way in six months, in a year, in 18 months. So a little bit of that crystal ball, a little bit of, you know, behind the scenes on the roadmap.

But, um, I wouldn't change anything I don't think. All right. I think that was a request.

Are you gonna share more in the roadmap and let people know what's coming down the pike a little more aggressively? We did. We did.

But, but I'm always there to do this more, right? And, uh, at the time permits, I can, I can throw some of the things which are coming in the pipeline, but we had a packed roadmap session, two of them, all of them with standing room only about what's coming in and where we are taking this across the whole automation engine. And the thing says, people are realizing is what Steve was saying, is when we talk about automation, we are not talking just about rpa, we are talking about integrations, document intelligence, um, um, N R P A plus, we are giving them a tool to manage all four of them in a way which as Steve's kind of increases his team width and have a more optimal use of his time and, and decentralize or centralize whatever model he decides to implement at U S D A.

So that's the another sort of areas of investments we are doing other than the fact that gen AI And even low tech, just a simple api, right? Like that's still a great tool. Yes.

And, and it's in the toolbox. And I think, you know, just as many options to accomplish, to torture a metaphor to skin the cat as many ways as possible, I think is, is never a bad thing. It's never a about it.

And then to that point, one more, just a quick plug here. We do integrations, as you know, we have 180 plus spokes as you know, our, but we also have something called custom spokes framework, meaning customers can write to endpoints With our upcoming Vancouver release, Michael, we are launching something called spoke generator. Any endpoint with an API specification, you point to it, it'll create a spoke for you in under two minutes.

And that would make the integration much more easier for people like Steve and his team. So that's what we are doing. All right, folks, you heard it here before you roll out that duct tape and start getting the bailing wire together, they've stitched something together.

Check out the product roadmap. Hey guys, thank You. Pleasure for being here.

Bye. Thanks so much. All Right.

And we'll be back in a minute.